[pull] main from mozilla:main #5

pull · 2024-05-07T06:57:49Z

See Commits and Changes for more details.

Can you help keep this open source service alive? 💖 Please sponsor : )

@mxinden

* fix: Send ACKs during handshake after they are lost This is not yet complete. It works for the first RTX of a lost server Initial, which now contains the same data as the original transmission. A third RTX only contains pings. Something is off about our probe logic. Will continue investigating, but want to see what the bench and QNS impacts of this change are. * Improvements * Remove debug output * Suggestion from @mxinden * doc comment * Check if `L1`/`C1` improve if this is also done in the Handshake space * Update comment * ACK more aggressively during the handshake * Undo * Update neqo-transport/src/tracking.rs Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/recovery/mod.rs Signed-off-by: Lars Eggert <[email protected]> --------- Signed-off-by: Lars Eggert <[email protected]>

* fix: qlog `reference_time` should be in msec https://datatracker.ietf.org/doc/html/draft-ietf-quic-qlog-main-schema#section-7.1 * Update neqo-common/src/qlog.rs Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Undo --------- Signed-off-by: Lars Eggert <[email protected]> Co-authored-by: Max Inden <[email protected]>

Bumps [docker/build-push-action](https://github.com/docker/build-push-action) from 6.6.1 to 6.7.0. - [Release notes](https://github.com/docker/build-push-action/releases) - [Commits](docker/build-push-action@16ebe77...5cd11c3) --- updated-dependencies: - dependency-name: docker/build-push-action dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

@martinthomson

* fix: Don't process timers during Closing or Draining This doesn't seem to cause any packets to go out, but is a bit wasteful (and also confused me when I saw timers fire in the logs...) * Suggestion from @martinthomson

* test: Add a test case for 0-RTT loss This currently fails, because we have a bug where the server does not accept retransmitted 0-RTT packets when it should. * Fixed * Cleanup * Keep the larger anti-replay window to just one test * Update neqo-transport/src/connection/tests/zerortt.rs Co-authored-by: Martin Thomson <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/connection/tests/zerortt.rs Co-authored-by: Martin Thomson <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/connection/tests/zerortt.rs Co-authored-by: Martin Thomson <[email protected]> Signed-off-by: Lars Eggert <[email protected]> --------- Signed-off-by: Lars Eggert <[email protected]> Co-authored-by: Martin Thomson <[email protected]>

* refactor: drop time crate in favor of std::time::SystemTime * use as_secs_f64 * don't panic, but return None

PMTUD probes inflate what we sent in 1-RTT, causing QNS to fail the test. See ptrd/kwik#46 (comment)

* fix: Lower log level for "Closing, not processing other timers" It's too noisy. * More

@mxinden

* test: Check for RTX of lost 0-RTT data on reject This is currently failing. * Fix merge * Attempt to RTX requests when 0-RTT was rejected Look OK, @mxinden? Let's see what the CI says. * Reinit on `ZeroRttRejected` * Remove test * Fix * Reinit files * Suggestions from @mxinden --------- Signed-off-by: Lars Eggert <[email protected]>

s2n-quic doesn't send a resumption token for the second connection in the QNS `resumption` test, and this caused us to fail.

* ci: Delete all QNS logs >10MB before artifact upload Because these artifacts are huge and those logs are not useful for debugging. * 5M * Leave a note

Signed-off-by: Lars Eggert <[email protected]>

Add test coverage for validation failing on `Ect1`. https://github.com/mozilla/neqo/blob/15cae9be1529e327e575fb7c799acc7521691337/neqo-transport/src/ecn.rs#L244-L246

* chore: Switch to the `windows` crate The `winapi` isn't being updated anymore, and the `windows` crate is a more modern alternative maintained by Microsoft. * Fix

) Previously `send_something_paced_with_modifier` would only apply the `modifier` if `sender.process_output` returned a `Output::Datagram` right away, but not when `sender.process_ouput` returned a datagram after returning a `Output::Callback`. With this commit, `modifier` is always applied.

It's been having weird failures.

* feat(transport): add ecn metrics to Stats * clippy * Document NotEct 0 * clippy * fix doc links

…ns (#2084) Bumps martenseemann/quic-network-simulator-endpoint from `5114afe` to `5177120`. --- updated-dependencies: - dependency-name: martenseemann/quic-network-simulator-endpoint dependency-type: direct:production ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* build(deps): bump actions/upload-artifact from 4.3.6 to 4.4.0 Bumps [actions/upload-artifact](https://github.com/actions/upload-artifact) from 4.3.6 to 4.4.0. - [Release notes](https://github.com/actions/upload-artifact/releases) - [Commits](actions/upload-artifact@834a144...5076954) --- updated-dependencies: - dependency-name: actions/upload-artifact dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> * Add actions --------- Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Lars Eggert <[email protected]>

* fix: Handle migrating to a path that is an ECN blackhole * Fix tests * Update neqo-transport/src/connection/tests/ecn.rs Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/path.rs Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/connection/tests/ecn.rs Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/path.rs Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/connection/tests/ecn.rs Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/connection/tests/ecn.rs Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Update neqo-transport/src/connection/tests/ecn.rs Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> * Fixes --------- Signed-off-by: Lars Eggert <[email protected]> Co-authored-by: Max Inden <[email protected]>

Save some trees

* ci: Try and speed up QNS docker build Via https://github.com/LukeMathWalker/cargo-chef * Update qns/Dockerfile Co-authored-by: Max Inden <[email protected]> Signed-off-by: Lars Eggert <[email protected]> --------- Signed-off-by: Lars Eggert <[email protected]> Co-authored-by: Max Inden <[email protected]>

* chore: Trim unused features Courtesy of https://github.com/ToBinio/cargo-features-manager. * More * Again * More * Fixes * Needed? * Again * Do not encode long RTT guesses in resumption tokens * Oops

* ci: Remove some things from `env` 1. `CARGO_TERM_COLOR: always` is always set by `dtolnay/rust-toolchain` 2. `RUST_BACKTRACE` is better set where needed, rather than everywhere. * Set `RUST_BACKTRACE: 1` on test steps

I keep finding that I need to flip a PR from "draft" to "ready for review" prematurely, just to get the CI to run benchmarks.

Signed-off-by: Lars Eggert <[email protected]>

Bumps lukemathwalker/cargo-chef from `ee4eb4c` to `d14930e`. --- updated-dependencies: - dependency-name: lukemathwalker/cargo-chef dependency-type: direct:production ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* build(deps): bump actions/upload-artifact from 4.5.0 to 4.6.0 Bumps [actions/upload-artifact](https://github.com/actions/upload-artifact) from 4.5.0 to 4.6.0. - [Release notes](https://github.com/actions/upload-artifact/releases) - [Commits](actions/upload-artifact@6f51ac0...65c4c4a) --- updated-dependencies: - dependency-name: actions/upload-artifact dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> * Add actions --------- Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Lars Eggert <[email protected]>

Bumps [vmactions/freebsd-vm](https://github.com/vmactions/freebsd-vm) from 1.1.5 to 1.1.8. - [Release notes](https://github.com/vmactions/freebsd-vm/releases) - [Commits](vmactions/freebsd-vm@debf37c...848dac7) --- updated-dependencies: - dependency-name: vmactions/freebsd-vm dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

There is no single-packet version of `on_packets_acked`. Signed-off-by: Max Inden <[email protected]> Co-authored-by: Lars Eggert <[email protected]>

Bumps [vmactions/solaris-vm](https://github.com/vmactions/solaris-vm) from 1.1.0 to 1.1.1. - [Release notes](https://github.com/vmactions/solaris-vm/releases) - [Commits](vmactions/solaris-vm@a89b943...cc8f82f) --- updated-dependencies: - dependency-name: vmactions/solaris-vm dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

From 20 to 30 min. With the latest Ubuntu runner, the C compile takes longer :-(

Bumps [docker/setup-qemu-action](https://github.com/docker/setup-qemu-action) from 3.2.0 to 3.3.0. - [Release notes](https://github.com/docker/setup-qemu-action/releases) - [Commits](docker/setup-qemu-action@49b3bc8...53851d1) --- updated-dependencies: - dependency-name: docker/setup-qemu-action dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Bumps [vmactions/openbsd-vm](https://github.com/vmactions/openbsd-vm) from 1.1.4 to 1.1.6. - [Release notes](https://github.com/vmactions/openbsd-vm/releases) - [Commits](vmactions/openbsd-vm@0cfe06e...7ac70b6) --- updated-dependencies: - dependency-name: vmactions/openbsd-vm dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* chore: `cargo clippy --fix` For new nightly lints. * fmt

On the new `ubuntu-latest` runner, the C compiler is slower.

* docs(transport/cubic): add references to RFC 8312 * Use htmlized * Clippy * Clippy

It fails if the runtime exceeds 10 seconds, which happens sometimes in CI, e.g., https://github.com/mozilla/neqo/actions/runs/12772087190/job/35600809738?pr=2295#step:7:231

* ci: Revamp the QNS CI action timeout It seems that the `TIMEOUT` env variable applies to all steps? * Simpler

#2269 made a "stream control or data frame for a stream that is ours to initiate but we haven't yet" illegal. ``` rust let existed = !stream_id.is_remote_initiated(self.role) && self.local_stream_limits[stream_id.stream_type()].used() > stream_id.index(); ``` Problem is, when `is_remote_initiated` is `true`, then `existed` is `false`, which leads to a false positive for remote initiated streams.

* ci: pin nightly * Trigger CI

This reverts commit e18658d.

This reverts commit c3ecaa6.

Bumps lukemathwalker/cargo-chef from `d14930e` to `9d68a1b`. --- updated-dependencies: - dependency-name: lukemathwalker/cargo-chef dependency-type: direct:production ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Bumps [docker/build-push-action](https://github.com/docker/build-push-action) from 6.10.0 to 6.12.0. - [Release notes](https://github.com/docker/build-push-action/releases) - [Commits](docker/build-push-action@48aba3b...67a2d40) --- updated-dependencies: - dependency-name: docker/build-push-action dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Bumps [codecov/codecov-action](https://github.com/codecov/codecov-action) from 5.0.7 to 5.1.2. - [Release notes](https://github.com/codecov/codecov-action/releases) - [Changelog](https://github.com/codecov/codecov-action/blob/main/CHANGELOG.md) - [Commits](codecov/codecov-action@v5.0.7...1e68e06) --- updated-dependencies: - dependency-name: codecov/codecov-action dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Bumps [vmactions/netbsd-vm](https://github.com/vmactions/netbsd-vm) from 1.1.4 to 1.1.8. - [Release notes](https://github.com/vmactions/netbsd-vm/releases) - [Commits](vmactions/netbsd-vm@7c9086f...46a58bb) --- updated-dependencies: - dependency-name: vmactions/netbsd-vm dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

) * deps(common/log): drop env_logger's auto-color and regex features #2291 added additional formatting to our log output via `env_logger`'s `auto-color` feature. The `auto-color` feature adds the `is-terminal` crate dependency. `is-terminal` depends on `hermit-abi` `v0.4`. Pulling latest Neqo `v0.12.0` into mozilla-central adds `is-terminal` and `hermit-abi` `v0.4.0` to the Firefox dependency tree. In order to keep our dependency footprint low, I suggest not enabling the two `env-logger` features. * chore: bump version to v0.12.1

Until we address the `FIXME`s.

pull bot added the ⤵️ pull label May 7, 2024

larseggert and others added 29 commits August 13, 2024 12:33

Refactor NSS initialization (#2050)

165fff6

Add a tiny bit of hygiene to this macro (#2060)

a9f707f

fix: Don't process timers during Closing or Draining (#2061)

318ec99

* fix: Don't process timers during Closing or Draining This doesn't seem to cause any packets to go out, but is a bit wasteful (and also confused me when I saw timers fire in the logs...) * Suggestion from @martinthomson

refactor: drop time crate in favor of std::time::SystemTime (#2059)

c35028a

* refactor: drop time crate in favor of std::time::SystemTime * use as_secs_f64 * don't panic, but return None

fix: Turn off PMTUD during QNS zerortt test (#2064)

836ffc6

PMTUD probes inflate what we sent in 1-RTT, causing QNS to fail the test. See ptrd/kwik#46 (comment)

chore: Lower some log levels (#2065)

90d1585

* fix: Lower log level for "Closing, not processing other timers" It's too noisy. * More

fix: Don't stop handler if there is no resumption token (#2067)

43ce605

s2n-quic doesn't send a resumption token for the second connection in the QNS `resumption` test, and this caused us to fail.

chore: Fix various issues that markdownlint found (#2066)

adf4722

ci: Delete all QNS logs >5MB before artifact upload (#2071)

15cae9b

* ci: Delete all QNS logs >10MB before artifact upload Because these artifacts are huge and those logs are not useful for debugging. * 5M * Leave a note

fix: Add missing space (#2074)

b4168d5

Signed-off-by: Lars Eggert <[email protected]>

test(transport/ecn): cover disable on remark (#2075)

910a7cd

Add test coverage for validation failing on `Ect1`. https://github.com/mozilla/neqo/blob/15cae9be1529e327e575fb7c799acc7521691337/neqo-transport/src/ecn.rs#L244-L246

chore: Switch to the windows crate (#2077)

a793d4a

* chore: Switch to the `windows` crate The `winapi` isn't being updated anymore, and the `windows` crate is a more modern alternative maintained by Microsoft. * Fix

ci: Don't cache nss

9e8d298

It's been having weird failures.

fix: Don't panic on Windows is GITHUB_WORKFLOW is not set (#2079)

22370e7

chore: Fix clippy nightly warnings (#2080)

7e8d22d

feat(transport): add ecn metrics to Stats (#2072)

6f8823b

* feat(transport): add ecn metrics to Stats * clippy * Document NotEct 0 * clippy * fix doc links

ci: Only build Firefox for non-draft PRs (#2085)

e41e639

Save some trees

chore: prepare v0.9.0 release (#2082)

28f60bd

chore: Trim unused features (#2090)

45a2bb1

* chore: Trim unused features Courtesy of https://github.com/ToBinio/cargo-features-manager. * More * Again * More * Fixes * Needed? * Again * Do not encode long RTT guesses in resumption tokens * Oops

larseggert and others added 30 commits January 13, 2025 13:59

ci: Remove some things from env (#2336)

a2c52f1

* ci: Remove some things from `env` 1. `CARGO_TERM_COLOR: always` is always set by `dtolnay/rust-toolchain` 2. `RUST_BACKTRACE` is better set where needed, rather than everywhere. * Set `RUST_BACKTRACE: 1` on test steps

ci: Run benchmarks on draft PRs (#2345)

f9a3a93

I keep finding that I need to flip a PR from "draft" to "ready for review" prematurely, just to get the CI to run benchmarks.

chore: We no longer need the tpids! macro (#2341)

9abde4e

Signed-off-by: Lars Eggert <[email protected]>

chore(transport/cc): remove obsolete comment (#2343)

a979459

There is no single-packet version of `on_packets_acked`. Signed-off-by: Max Inden <[email protected]> Co-authored-by: Lars Eggert <[email protected]>

ci: Lengthen QNS CI timout

81e0800

From 20 to 30 min. With the latest Ubuntu runner, the C compile takes longer :-(

chore: cargo clippy --fix (#2355)

4f138a4

* chore: `cargo clippy --fix` For new nightly lints. * fmt

ci: Lengthen QNS CI timout (#2356)

1a38b27

On the new `ubuntu-latest` runner, the C compiler is slower.

ci: Run cargo-deny during ci (#2340)

8bc926a

docs(transport/cubic): add references to RFC 8312 (#2344)

7bbf900

* docs(transport/cubic): add references to RFC 8312 * Use htmlized * Clippy * Clippy

ci: Fix the add_event_with_instant test (#2359)

7f8136e

It fails if the runtime exceeds 10 seconds, which happens sometimes in CI, e.g., https://github.com/mozilla/neqo/actions/runs/12772087190/job/35600809738?pr=2295#step:7:231

ci: Revamp the QNS CI action timeout (#2361)

b550010

* ci: Revamp the QNS CI action timeout It seems that the `TIMEOUT` env variable applies to all steps? * Simpler

ci: pin nightly (#2370)

e18658d

* ci: pin nightly * Trigger CI

Revert "ci: pin nightly (#2370)" (#2371)

a7b33a3

This reverts commit e18658d.

Revert "ci(sanitize): downgrade to ubuntu v22.04 (#2353)" (#2364)

b4143a1

This reverts commit c3ecaa6.

chore: prepare v0.12.0 (#2362)

a7e2fb0

ci(clippy): include benchmarks (#2372)

7218cff

deps: update mtu to v0.2.4 (#2379)

29f63d0

ci: Disable failing VM actions in CI (#2380)

f06c7ec

Until we address the `FIXME`s.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[pull] main from mozilla:main #5

[pull] main from mozilla:main #5

pull bot commented May 7, 2024 •

edited

Loading

[pull] main from mozilla:main #5

Are you sure you want to change the base?

[pull] main from mozilla:main #5

Conversation

pull bot commented May 7, 2024 • edited Loading

pull bot commented May 7, 2024 •

edited

Loading